Google multimodal search: how people find you with a photo and what Search Console shows

Google multimodal search means searching with an image instead of a typed query: someone takes a photo of a product, a plant or a sign, and Google shows them answers and websites. Since September 24, 2026, Search Console shows how many of these searches brought people to your site. The new search type filter covers Google Lens, Circle to Search on Android, uploaded images and image search in Chrome. The data only appears when your site actually receives traffic from such searches.

Below you will find how multimodal search differs from Google Images, where exactly you see it in Search Console, how to read the numbers and what to change on your site so people can find you with a photo at all. For the bigger picture of how AI is changing search, read our article on Google AI Mode and its impact on your website. How to measure traffic from chatbots is covered in our guide to AI traffic in GA4. If you want to prepare your site with us, take a look at our SEO optimisation service.

What Google multimodal search is and how it differs from Google Images

In multimodal search the image itself is the query, while in Google Images people type text and get images as results.

Multimodal means that search works with more than one kind of input, not just text. According to the announcement on the Google Search Central blog, the new data counts four ways of searching:

• Google Lens: the person points their phone camera at something around them or takes a photo of what interests them.
• Circle to Search on Android: they circle what they want to search for on the phone screen.
• Uploaded image: on google.com they upload a file from their computer or drag it into the search box.
• Image search in Chrome: they right-click an image on any website and let Google search by it.

According to Google's help page on searching with an image, after uploading an image people can also type text into the "Add to your search" box, such as a colour or a question. The results can then include an AI overview, results for objects in the image, similar images and websites with the same or a similar image. That last item is the interesting one for a business: if your site has your own photo of a product, it can appear among the results.

Search Console now distinguishes five search types in the Performance report. The difference matters, because each one counts something else:

Search type in Search Console	What it counts	Typical example
Web: text-based	Queries typed into the standard Google search bar	Someone types "ceramic coffee mug"
Web: multimodal	Searching with an image: Lens, Circle to Search, uploaded images, image search in Chrome	Someone photographs a mug at a friend's place and looks for where to buy it
Image	Results in the Google Images tab	Someone searches with text and browses images
Video	Results in the Videos tab	Someone looks for a video tutorial
News	The News tab and the news search results view	Someone follows news on a topic

Multimodal search in Search Console: where to find it step by step

You find multimodal search in the Performance report by switching the Search type filter to Web: multimodal; there is nothing to enable or install.

According to the announcement, Google has been rolling the feature out to sites worldwide since September 24, 2026 (as of October 2026):

1. Open the Performance report: in Search Console choose your property and, in the left menu, Performance, then Search results.
2. Switch the search type: above the chart click the Search type filter and choose Web: multimodal. The search type filter cannot be removed, only changed.
3. Check the metrics: turn on clicks, impressions, CTR and average position. The report shows them just as it does for regular search.
4. Find out which pages appear: below the chart switch the table to Pages. You will see which of your pages Google showed to people who searched with an image.
5. Compare with text search: in the Search type filter choose Compare and set Web: text-based against Web: multimodal. The Search Console help page on filtering and comparison states that you can only see several search types at once through a comparison, which is limited to two items.
6. Export the data: the Export button downloads both the chart and the table, so you can compare month by month in a spreadsheet.

The same filter is in the generative AI features report

The second place where you find the filter is the generative AI performance report. According to the Search Console help page for this report, it has been available for all websites since August 31, 2026 and shows impressions of your site in AI Overviews and AI Mode. Here, too, you can choose Web: text-based or Web: multimodal. The difference from the Performance report: this report shows impressions only, not clicks. We write more about getting into AI Overviews answers in our article Google AI Overviews and your SEO.

How to read the numbers from the multimodal filter

The most valuable part is not the total but the list of pages: it shows for which products or services people search for you with a photo.

When you first look at the new filter, keep a few things in mind:

• An empty report is not an error. Google's announcement says you will see the metrics if your site receives traffic from these searches. For a brochure website of a service business, the number can stay at zero for a long time.
• Multimodal is part of web search. The Search Console help describes the Web type as text-based or multimodal. So watch the ratio between these two parts and do not add them up with the other types.
• Do not confuse multimodal with Image. Image shows how people find you in the Google Images tab after typing text. Multimodal shows the opposite direction, searching with an image.
• The newest data still changes. Search Console marks preliminary data with a dotted line in the chart, and it can change within the next few hours.
• The history is still short. Google's announcement does not say how far back the data goes. So export it regularly and watch the trend.

In practice: if a specific product, category or service appears in the list of pages, it is a signal that people are searching for something you photographed. Check those pages first. Do they have your own sharp photos, a descriptive alt text and clear text next to the image? If not, that is where to start.

How to prepare your website for photo search: 7 steps

You need no special markup or new files; what matters is whether Google finds your images, indexes them and understands them from the surrounding text.

In its documentation on AI features and your website, Google writes that there are no additional technical requirements and no special structured data needed to appear in AI Overviews and AI Mode. Among its recommendations it lists supporting your text with high-quality images and keeping your Merchant Center and Business Profile information up to date. The concrete steps follow Google's image SEO best practices:

1. Embed images as an HTML img element. Google finds the image in the src attribute, even inside a picture element. Google does not index images set as a CSS background.
2. Write descriptive alt text. Google calls it the most important image attribute and uses it together with computer vision and the page content. Describe what is in the photo, without keyword stuffing.
3. Give files meaningful names. A name like black-ceramic-mug.jpg gives Google at least a light clue, IMG00023.JPG gives none. For language versions, translate the file names too.
4. Place the image near relevant text. Google reads the subject of an image from the page content, including captions. A product photo belongs next to its description, not in a gallery without text.
5. Add images to your sitemap. An image sitemap helps Google find images it might not otherwise discover, for example ones your site loads with JavaScript.
6. Use supported formats and fast loading. Google Search supports BMP, GIF, JPEG, PNG, WebP, SVG and AVIF. Sharp photos attract more, but images are often the largest part of a page, so resize them and keep a fallback src attribute for responsive versions.
7. Specify the main image of the page. With og:image or the image property in structured data you tell Google which image represents the page. It should not be a logo or an image with text.

For speed, our article on how to check your website speed shows how to find images that are too heavy with PageSpeed Insights.

Multimodal search in practice: what we see on business websites

In our experience, the biggest obstacles are product photos copied from the manufacturer, images hidden in a CSS background and empty or meaningless alt texts.

In audits of business websites and online shops we keep running into the same patterns:

• The same photo as the competition. Smaller online shops often reuse product photos from their supplier. When someone searches with a photo, Google shows websites with the same or a similar image, and you are just one of many. Your own photo sets you apart.
• Hero slideshows in CSS. On template-based websites the main images are often inserted as a background. It looks good, but Google does not index them.
• Alt text as the file name. Values like "DSC_0412" or the same alt on every photo in a gallery tell Google nothing.
• Outdated data outside the website. For local businesses and shops, the Business Profile or the product feed often does not match what is on the website.

Multimodal search makes the most sense where customers see a thing before they know its name: furniture, clothing, spare parts, plants, food, but also craftwork or building materials. If you sell something people see at a friend's place or in the street, the new filter shows you for the first time whether Google displays you for such a search.

How we handle it at Denva: we review the website in Search Console and with a crawl of the whole structure, and you get a list of issues ranked by their impact on rankings. Typical fixes include faster image loading, a clean sitemap and structured data via Rank Math. We set up measurement in your own Search Console and GA4, and the access stays with you.

When you can do it yourself and when to bring us in

You can handle it yourself if:

• you have access to Search Console and only want to find out whether people search for you with a photo,
• your site has dozens of pages and you can add alt texts or file names in the admin,
• you can take photos of your products or projects yourself in good light.

It is worth bringing us in if:

• you run an online shop with hundreds of products and fixing images by hand makes no sense,
• your main images are inserted via CSS or loaded with JavaScript and you do not know whether Google sees them,
• images slow your site down and you want to speed them up without losing quality.

Conclusion

Google multimodal search is no longer just a mobile curiosity. Since September 24, 2026, Search Console shows it as a separate Web: multimodal type, both in the Performance report and in the generative AI features report. Check which pages appear there and start with them: your own photos, images in an img element, descriptive alt text, text next to the image and fast loading. If you want your website reviewed properly, get in touch through a no-obligation consultation.

FAQ

Frequently asked questions.

What is Google multimodal search?

It is searching with an image instead of typed text. Search Console counts Google Lens, Circle to Search on Android, uploaded images and image search in the Chrome browser as multimodal search.

Where do I find multimodal search in Search Console?

In the Performance, Search results report after switching the Search type filter to Web: multimodal. The generative AI performance report has the same filter.

Since when does Search Console show multimodal search?

Google announced the feature on September 24, 2026 and has been rolling it out to sites worldwide since that day. You will see the data if your site receives traffic from such searches.

What is the difference between the Web: multimodal and Image types?

Image counts results in the Google Images tab, where people type text and get images. Web: multimodal counts the opposite direction, when the image itself is the query, for example a phone photo.

Do I need special markup or files for multimodal search?

No. Google states that AI features need no additional technical requirements and no special structured data. What matters are the image SEO basics: the img element, descriptive alt text, text next to the image, a sitemap and fast loading.